BIOTROPICA 32(1): 70–79 2000

Insights into the Witheringia solanacea (Solanaceae) Complex in Costa Rica. I. Breeding Systems and Crossing Studies1

Lynn Bohs2 Department of Botany, Duke University, Durham, North Carolina 27708, U.S.A.

ABSTRACT The Witheringia solanacea complex consists of three species, W. asterotricha, W. meiantha, and W. solanacea, native to Central and South America. The three taxa are morphologically similar, and their distinctions and relationships have been the subject of taxonomic controversy. To investigate breeding systems and potential for hybridization among the taxa of the complex, two Costa Rican accessions per species were used in a crossing program. All plants were self- incompatible except for one accession of W. solanacea. Hybrid plants resulted from all crosses among accessions of W. asterotricha and W. solanacea. Most crosses were unsuccessful using W. meiantha in combination with either of the other two taxa. It is suggested that W. meiantha and W. solanacea be recognized as separate taxa, but that W. asterotricha be considered a synonym of W. solanacea.

RESUMEN El complejo Witheringia solanacea consiste de tres especies de Centro y Surame´rica, W. asterotricha, W. meiantha, y W. solanacea. Las tres especies son morfolo´gicamente parecidas entre sı´ y han sido fuente de muchas controversias tax- ono´micas. Para investigar el sistema reproductivo y el potencial para hibridizacio´n entre las especies del complejo, se utilizaron seis colecciones de Costa Rica (dos por especie) en un programa de cruzamiento en invernadero. Todas las plantas eran auto-incompatibles salvo una coleccio´n de W. solanacea. Se formaron hı´bridos en todas las combinaciones entre las colecciones de W. asterotricha y W. solanacea. La mayorı´a de intentos entre W. meiantha y las otras dos especies fallaron. Se sugiera un esquema taxono´mico para el complejo que reconoce dos especies, W. meiantha y W. solanacea, con W. asterotricha como sino´nimo de W. solanacea.

Key words: breeding systems; Costa Rica; crossing studies; self-incompatibility; Solanaceae; Witheringia.

THE GENUS WITHERINGIA L’H ER.(SOLANACEAE) IN- ingia may be more closely related to the physaloid CLUDES ca 20 species of Neotropical herbs, shrubs, genera [Physalis L., Margaranthus Schlecht., Leu- and small trees. The center of diversity for the ge- cophysalis Rydb., Chamaesaracha (A. Gray) Benth.] nus is Costa Rica, where over half the species occur. and that the genus may not be monophyletic as Witheringia has been placed in subfamily Solan- currently circumscribed (Olmstead et al. 2000; oideae, tribe Solaneae, by virtue of its flattened Bohs, pers. obs.). seeds with curved embryos, abundant endosperm, The genus Witheringia is distributed from basal filament insertion, and valvate corolla aesti- Mexico to Bolivia, with the majority of species oc- vation (Hunziker 1979, D’Arcy 1991). Witheringia curring in Mexico, Costa Rica, and Panama. Hun- has been allied with the genera Acnistus Schott, ziker’s (1969) revision treated 15 taxa, including 3 Athenaea Sendtn., Aureliana Sendtn., Brachistus species of section Brachistus. He pointed out that Miers, Capsicum L., Cuatresia Hunz., Dunalia 3 species, W. asterotricha (Standl.) Hunz., W. H.B.K., Iochroma Benth., Saracha Ruiz & Pav., and meiantha (Donn. Sm.) Hunz., and W. solanacea Vassobia Rusby in traditional classification schemes L’Her., form a tightly related group termed the W. (Hunziker 1984, 1987) and often has been consid- solanacea complex. All are herbs or shrubs with ered as being closely related to Capsicum because truncate calyces and a ring of hairs inside the co- of similarities in flower and fruit characters. Recent rolla tube, and with corolla lobes much longer than molecular studies, however, indicate that Wither- the tube. These species have very short (Ͻ5 mm) peduncles or lack them altogether, so that the flow- 1 Received 15 January 1998; revision accepted 7 Novem- ers are arranged in fascicles in branch forks or op- ber 1998. posite the paired leaves. Witheringia solanacea has 2 Current address: Department of Biology, University of Utah, Salt Lake City, Utah 84112, U.S.A.; e-mail: the broadest geographical distribution, ranging [email protected]. from southern Mexico through Central America

70 Crossing Studies in Witheringia 71 and the Antilles to South America. It is an extreme- ductive interaction among the taxa is at least the- ly variable species, but is distinguished by its pu- oretically possible. bescence of unbranched or sparsely branched hairs In addition to the crossing results, a brief key and often tetramerous flowers. Hunziker (1969) to the three entities is included at the conclusion listed one variety, W. solanacea var. silvigaudens of the paper. A complete systematic treatment of (Standl. & Williams) Hunz., endemic to Dept. the genus Witheringia is being prepared by M. Sou- Moraza´n, Honduras. This variety is glabrous, has sa-Pen˜a of the University of Connecticut. pentamerous corollas, and longer, narrower leaves than typical W. solanacea. Witheringia asterotricha MORPHOLOGY is closely related to W. solanacea, and according to Hunziker (1969), probably should be considered a The three taxa of the complex are all erect herbs, variety of the latter species. Witheringia asterotricha, weakly woody shrubs, or small trees reaching sev- thus far known only from Costa Rica and Panama, eral meters in height. The leaves are entire, elliptic, is morphologically distinguishable from W. solana- unequal in size, and paired at the nodes (Fig. 1A). cea by its dense pubescence of dendritically All three taxa lack obvious peduncles, and as a re- branched hairs and pentamerous flowers. Wither- sult, flowers are clustered in axillary fascicles (Fig. ingia meiantha ranges from southern Mexico to 1). The corollas are stellate and light yellowish to western Panama. Plants of W. meiantha are gla- cream in color, with those of W. meiantha smaller brous, with small pentamerous flowers and fewer- and deeper yellow than those of W. asterotricha and flowered inflorescences than W. solanacea or W. as- W. solanacea. The latter two taxa often have green- terotricha. ish maculations in the corolla throat, and the co- Unlike Hunziker (1969), D’Arcy (1973) rec- rolla lobes are longer and less reflexed than in W. ognized only two species in the complex. He main- meiantha. The tapered anthers are connivent tained W. asterotricha as distinct, but indicated that around the style, and begin to dehisce by large ter- minal pores that eventually open into longitudinal hybrids with W. solanacea may occur when the two slits (Fig. 1B). After pollination, the corolla and species come into contact. D’Arcy (1973) consid- attached stamens wither and fall, leaving only the ered W. meiantha a synonym of W. solanacea, com- calyx and gynoecium. As fruits develop, the pen- menting that the reduced pubescence and slightly dulous flowering pedicels curve upward until the larger calyces characterizing W. meiantha were like- ripe fruits are held erect (Fig. 1A, C, D). At ma- ly minor morphological variants of the widespread turity, all three taxa have red, shiny, juicy, many- W. solanacea. seeded fruits. Witheringia meiantha has glabrous Nee (1986), treating only the taxa from Vera- fruits, those of W. solanacea are glabrous to very cruz, Mexico, recognized W. meiantha as distinct sparsely pubescent, and those of W. asterotricha are from W. solanacea and considered W. solanacea var. moderately to rather densely pubescent. Fruit color silvigaudens a synonym of W. meiantha. According changes from green to orange–red very quickly, of- to Nee (1986), W. meiantha can be distinguished ten in a single day. These attributes are character- from W. solanacea by its larger calyx and seeds, pen- istic of dispersal by birds, or ornithochory (van der tamerous flowers, glabrous leaves, and few-flowered Pijl 1982). Bird dispersal of Witheringia fruits has inflorescences. been documented at La Selva and Monteverde, To gain further insight into the relationships of Costa Rica, by Wheelwright et al. (1984), Murray these taxa, I carried out a crossing study in the (1987, 1988), Loiselle and Blake (1990), Blake and greenhouse to determine if genetic isolating mech- Loiselle (1992), and Murray et al. (1994). anisms were operating among members of the W. solanacea complex. It was also of interest to ex- amine breeding systems of the three taxa, as no STUDY SITES AND HABITATS information of this kind is available for Withering- One accession of W. asterotricha and both acces- ia, and the distribution of self-incompatibility (SI) sions of W. meiantha used in the crossing study among genera of the Solaneae is not well known. were collected at La Selva Biological Station near Greenhouse plants were grown from seed collec- the confluence of the Sarapiquı´ and Puerto Viejo tions (accessions) of three sites in Costa Rica. Costa Rivers in Heredia Province (Table 1). Elevations at Rica, and perhaps Panama, are the most likely areas the station range from ca 35 to 140 m. La Selva of sympatry or near sympatry for the three species, has an annual mean temperature of 26ЊCandan and therefore represent localities in which repro- annual mean rainfall of ca 4000 mm. Although 72 Bohs

FIGURE 1. A. Witheringia asterotricha branch and inflorescence with flowers and young fruits. B. Flower of W. solanacea. C. Fruits of W. asterotricha. D. Flowers and fruits of W. meiantha.

rain falls at La Selva in all months of the year, there (Route 9) near the Catarata de La Paz at an elevation is a relatively wet season from late April to the of ca 1350 to 1400 m. Climatic data are not available beginning of August, with a smaller wet period for the site, but the average temperature is lower than from November through January. Detailed descrip- at La Selva, and the annual rainfall is probably greater. tions of the site can be found in McDade et al. One accession of W. solanacea was collected here at a (1994). Witheringia asterotricha and W. meiantha disturbed site by the roadside in full sun. Other With- are fairly common on the station property. With- eringia species found in the immediate vicinity in- eringia asterotricha is found in open, sunny areas of cluded W. meiantha, W. cuneata (Standl.) Hunz., W. secondary vegetation. Witheringia meiantha grows maculata (Standl. & Morton) Hunz., and W. fusco- in more shady habitats in light gaps of the forest violacea (Cufod.) Hunz. canopy, such as the margins of treefall gaps, trails, The Las Cruces Biological Station is located and streams. The two species were not observed near San Vito in the Coto Brus region of Puntar- growing together at La Selva. enas Province in southeastern Costa Rica. Eleva- One accession of W. asterotricha used in this tions range from ca 1100 to 1500 m. Las Cruces study was collected near Cariari, ca 32 km SW of is more seasonal than La Selva, with the dry season La Selva (Table 1). The climate in this area is sim- extending from January to March, when there is ilar to that of La Selva (J. Denslow, pers. comm.). little or no rain. Annual rainfall averages ca 4000 Witheringia asterotricha was found in full sun by mm, and the average daytime temperature is be- the roadside. No other Witheringia species were ob- tween 21 and 26ЊC (Organization for Tropical served in the immediate vicinity. Studies, pers. comm.). One accession of W. solan- Another field site was located above La Selva on acea was collected at Las Cruces from disturbed the road between Puerto Viejo and Vara Blanca areas in full sun along the access road and trails. Crossing Studies in Witheringia 73

TABLE 1. Sources of Witheringia plants used in this study. Seeds were collected from a single plant. Number of plants used in crosses is listed under each accession.

Witheringia asterotricha. Bohs 2377. Costa Rica: Prov. Limo´n, Portico logging project near Cariari, ca 10Њ20ЈN, 83Њ45ЈW, ca 50 m elev. Seed accession no. 91-13; 5 plants. Witheringia asterotricha. Bohs 3007. Costa Rica: Prov. Heredia, La Selva Biological Station, 10Њ26ЈN, 83Њ59ЈW, Sen- dero La Chanchera (SCH) near junction with Camino Experimental Norte (CEN) (approx. grid coordinates 600 ϫ 100), ca 50–60 m elev. Seed accession no. 90-3; 3 plants. Witheringia meiantha. Bohs 3015. Costa Rica: Prov. Heredia, La Selva Biological Station, 10Њ26ЈN, 83Њ59ЈW, Sende- ro Oriental (SOR) just past waterfall (approx. grid coordinates 600 ϫ 1300), ca 40–60 m elev. Seed accession no. 90-7; 5 plants. Witheringia meiantha. Bohs 2387. Costa Rica: Prov. Heredia, La Selva Biological Station, 10Њ26ЈN, 83Њ59ЈW, Sende- ro Sura´ (SUR) at entrance to arboretum (approx. grid coordinates 800 ϫ 600), ca 40–60 m elev. Seed accession no. 91-22; 2 plants. Witheringia solanacea. Bohs 2416. Costa Rica: Prov. Alajuela, Rio La Paz Pequen˜a, ca 500 m N of Catarata de La Paz next to road between Puerto Viejo and Vara Blanca, ca 10Њ13ЈN, 84Њ10ЈW, ca 1350–1400 m elev. Seed accession no. 92-8; 10 plants. Witheringia solanacea. Bohs 2427. Costa Rica: Prov. Puntarenas, vicinity of Las Cruces Biological Station near San Vito, margin of road between station and Agua Buena, ca 8Њ50ЈN, 83Њ05ЈW, ca 1100 m elev. Seed accession no. 92-13; 10 plants.

Witheringia meiantha also was found at Las Cruces, plants, but only a subset of the F1 plants reached where it was uncommon in light gaps of the pri- flowering maturity. mary forest. Pollen tube growth was observed using the technique of Martin (1959) with the following modifications. Flowers were harvested and fixed MATERIALS AND METHODS 48–72 hours after pollination. Gynoecia were then Њ Seeds were gathered from ripe fruits at the study rinsed and cleared in 0.8 N NaOH at 60 C for 1 sites described above and plants were grown in pol- to 2 hours, rinsed briefly, and stained. At least three linator-free greenhouses at the University of Utah. flowers were examined per crossing combination. Sources of plants and voucher information are giv- Chromosome observations were made from en in Table 1. fresh flower buds gathered from greenhouse-grown For the artificial crosses, pollen was tapped plants. Anthers were macerated in a 1.0 percent onto a clean glass slide and rubbed on the stigma acetocarmine solution or in LP orcein (W. Bloom, pers. comm.), a cover slip was added, and the prep- of the maternal parent. Plants suspected or deter- aration was squashed. Herbarium vouchers were mined to be self-compatible were emasculated in deposited at the National Museum in San Jose´, the bud before pollination. Numbers of crosses per- Costa Rica, and at the Garrett Herbarium at the formed, number of fruits set, and number of seeds University of Utah, Salt Lake City. Type specimens per fruit are listed in Table 2. ‘‘Sib crosses’’ refer to of Acnistus lehmanni Damm., Capsicum asterotri- crosses involving individuals from a single seed ac- chum Standl., C. costaricense Standl. & Morton, C. cession, whereas ‘‘outcrosses’’ refer to crosses be- isothrix Standl., C. multiflorum Standl. & Morton, tween individuals from different seed accessions. In C. silvigaudens Standl. & Williams, C. stenophyllum crosses yielding fruits, the fruit size, color, and seed Morton & Standl., C. tetramerum Standl. & Mor- number and appearance were recorded. Seeds sub- ton, Sicklera tetrandra A. Braun & Bouche´, and sequently were planted in the greenhouse to deter- Solanum panamense van Heurck & Muell.-Arg. mine their viability and to assess the vitality of the were examined to ensure that my taxonomic con- F plants. 1 cepts were congruent with those of earlier authors. Pollen viability was determined by shaking These names and types include a majority of the grains onto a clean glass slide and staining in ani- nomenclatural and taxonomic synonyms that have line blue-lactophenol (Hauser & Morrison 1964). been used for the taxa of the W. solanacea complex. Pollen grains were allowed to stain for at least one hour before scoring. The first 300 grains encoun- tered were scored, and unshriveled grains staining RESULTS blue in the preparation were presumed to be viable. CHROMOSOME NUMBERS.—Meiotic squashes from Pollen stainability was measured in all parent accessions of all three taxa (W. asterotricha: Bohs 74 Bohs 0 0 0 0 2 22 18 21 14 26 20 24 14 15 6–64 (38) 13–54 (36) 18–77 (48) 32–104 (75) . For seed viability, 6 0 0 0 0 25 23 19 25 24 21 11 20 16 10–29 (19) 29–63 (46) 16–87 (50) 12–86 (56) 9 3 2 2 20 24 22 14 23 12 24 19 19 (19) 8–24 (14) 2–36 (22) 45–48 (46) 16–32 (24) 40–51 (46) Male parent 0 0 0 0 3 20 28 30 12 22 11 34 11 23 2–47 (16) 5–59 (29) 7–37 (22) 10–34 (21) 0 0 0 0 5 19 18 37 11 50 21 25 14 18 7–78 (37) 7–85 (39) 14–28 (24) 19–77 (54) 0 0 0 0 18 16 18 17 22 29 30 17 23 19 ast 3007 ast 2377 mei 3015 mei 2387 sol 2427 sol 2416 6–93 (50) 32–69 (55) 30–58 (41) 30–76 (56) Taxon abbreviations follow Figures 2 and 3. Intertaxon crosses are outcrosses or sib crosses unless noted otherwise. All flowers 0 0 0 0 0 0 0 0 0 0 6 20 30 26 23 71 31 Witheringia. 33–75 (59) a No. fruits No. seeds/fruit No. fruits No. seeds/fruit No. fruits No. seeds/fruit No. fruits No. seeds/fruit No. fruits No. seeds/fruit No. fruits No. seeds/fruit Intra- and intertaxon crosses in of sol 2416see were Figure emasculated 2 in and the text. bud before pollination. Number of seeds/fruit was based on visual examination; only full-sized seeds were counted ). ¯ x Range ( Female parent Selfed TABLE 2. ast 3007 No. pollinations a ast 2377 No. pollinations mei 3015 No. pollinations mei 2387 No. pollinations sol 2427 No. pollinations sol 2416 No. pollinations Crossing Studies in Witheringia 75

TABLE 3. Pollen stainability of Witheringia parents and hybrids. Abbreviations for accessions follow Figures 2 and 3.

No. flowers No. plants Mean percent (SD) Range (%) Parents ast 3007 7 4 90.7 (6.8) 83–99 ast 2377 6 3 93.9 (2.4) 90–97 mei 3015 25 5 80.9 (13.7) 27–94 mei 2387 10 2 94.3 (7.3) 77–99 sol 2427 9 8 84.3 (12.6) 54–97 sol 2416 6 3 83.7 (12.8) 68–99 Overall 63 25 86.1 (12.2) 27–99 Hybrids ast 3007 ϫ sol 2427 10 5 57.3 (5.2) 50–66 ast 2377 ϫ sol 2427 4 4 71.6 (5.9) 65–79 sol 2427 ϫ ast 3007 5 5 57.9 (10.5) 42–66 sol 2427 ϫ ast 2377 5 4 67.0 (6.0) 58–75 sol 2427 ϫ mei 2387 5 1 62.6 (2.8) 60–66 Overall 29 19 61.9 (8.0) 42–79

3007; W. meiantha: Bohs 3015; W. solanacea: Bohs BREEDING SYSTEM.—Both accessions of W. asterotri- 2427) had N ϭ 12 chromosomes. Chromosomes cha and W. meiantha were SI, as judged by the at meiotic metaphase ranged from ca 1.5 to 3.5 results of artificial pollinations and by pollen tube ␮m long. growth observations (Table 2; Figs. 2 and 3). No fruits were set by any plant upon self-pollination POLLEN FERTILITY.—Results of the pollen stainabil- (Table 2; Fig. 2), although fruits and seeds fre- ity tests are given in Table 3. In general, all parent quently resulted from sib crosses or outcrosses. Sev- plants had high percentages of stainable pollen eral W. asterotricha and W. solanacea plants spon- throughout the study. Mean pollen stainability in taneously developed small fruits that occasionally the hybrids was consistently lower than in the pa- contained a few seeds. Numerous pollen tubes grew rental accessions (Table 3). into the ovary and around the ovules in sib crosses and outcrosses in W. asterotricha and W. meiantha, but all tubes were inhibited in the style in self-

FIGURE 2. Results of crossing study. Taxon abbrevia- tions correspond to accessions listed in Tables 1 and 2; numbers 1 through 6 refer to same taxa in rows and columns. Minus sign ϭ no fruits produced; open circle FIGURE 3. Results of pollen tube observations. Taxon ϭ full-sized fruits, but seeds not viable; filled circle ϭ abbreviations as in Figure 2. Plus sign ϭ most pollen full-sized fruits with viable seeds that subsequently pro- tubes reach ovary and ovules; plus/minus sign ϭ afew ϭ duced F1 plants. Intraspecific pollinations fall along di- pollen tubes occasionally seen in ovary; minus sign all agonal line; space above line ϭ sib crosses or outcrossed, pollen tube growth arrested in style. Intraspecific polli- space below line ϭ selfed. nations are selfs. See Figure 2 for explanation of diagonal. 76 Bohs pollinations (Fig. 3). The site of inhibition of pol- costaricense Standl. & Morton), pointing to the len tube growth in selfed pollinations was in the likelihood of natural hybridization between W. as- upper style, consistent with the system of game- terotricha and W. solanacea. tophytic SI found in many other Solanaceae. All plants of accession 2427 of W. solanacea DISCUSSION appeared to be SI. In contrast, several individuals of accession 2416 set fruits and seeds after self- BREEDING SYSTEM.—All three Witheringia species pollination. Pollen tube observations confirmed were found to be SI, thus increasing to nine the that self pollen was able to germinate and grow number of solanaceous genera for which SI has down the styles into the ovaries in these plants. been reported (the others are Brugmansia, Brunfel- Other individuals of accession 2416 did not exhibit sia, Capsicum, Lycium, Nicotiana, Petunia, Physalis, this response, and instead behaved like SI plants. and Solanum [including Cyphomandra and Lycoper- sicon]) (Fryxell 1957; Heiser & Smith 1958; Lock- INTERSPECIFIC CROSSES.—Most crosses between W. wood 1973a, b; Plowman 1973; Whalen & An- meiantha and the other two Witheringia species derson 1981; Charlesworth 1985; Bohs 1991; were unsuccessful. All crosses failed when either ac- Preißel & Preißel 1991). Brunfelsia, Nicotiana, and cession of W. meiantha was used as the female par- Petunia belong to subfamily Cestroideae, which oc- ent in combination with W. asterotricha or W. so- cupies a basal position in the family according to lanacea pollen (Fig. 2). In these crosses, large num- phylogenetic analyses of molecular data (Olmstead bers of pollen grains germinated to produce pollen & Palmer 1992, Olmstead et al. 2000). Brugman- tubes, but all pollen tube growth was inhibited in sia, Capsicum, Lycium, Physalis, Solanum, and With- the style (Fig. 3). Occasional full-sized fruits re- eringia belong to the more derived subfamily So- sulted from pollinations of W. asterotricha and W. lanoideae. More data on the distribution of breed- solanacea with W. meiantha pollen, but the seeds ing systems in Solanaceae are needed to determine from these fruits were smaller than usual and failed whether SI is the ancestral condition within the to germinate. F1 hybrid plants were produced in family, or whether SI evolved from self-compatible only one interspecific crossing combination using (SC) ancestors (cf. Weller et al. 1995). Similarly, W. meiantha (sol 2427 X mei 2387), and this result better knowledge of breeding systems and phylo- was obtained with only one individual from each genetic relationships is needed to ascertain the basal accession. Hybrid plants were sparsely to moder- condition in subfamily Solanoideae and tribe So- ately pubescent with unbranched hairs. One plant laneae to which Witheringia belongs. Furthermore, flowered, and had 5-merous (rarely 4-merous) co- given the possibility that Witheringia itself may not rollas and the uniform corolla coloration of the W. be monophyletic, phylogenetic relationships and meiantha parent. Pollen stainability was somewhat breeding systems of Witheringia and its putative lower in the hybrid than in the parent plants (Table relatives should be examined in detail to determine 3). Pollen tube growth in crosses using W. meiantha the evolutionary directions and constraints on the pollen with W. asterotricha or W. solanacea was var- development of SI in this clade. iable, with most pollen tubes inhibited in the style; Plants of W. solanacea 2416 showed polymor- occasionally, however, a few grew to the ovules (Fig. phism in breeding system, with some individuals ex- 3). hibiting SI and others SC. Changes in breeding sys- In contrast, crosses between all accessions of W. tem, usually from SI to SC, have been reported for asterotricha and W. solanacea were successful in other solanaceous plants (Nettancourt 1977), and both directions (Fig. 2). The resulting hybrid have occurred many times over the course of angio- plants grew vigorously and flowered profusely. A sperm evolution (e.g., Jain 1976, Weller et al. 1995). subset of the hybrid plants was examined for pollen W. solanacea 2416 may represent the early stages of fertility, and the hybrids on average had lower pol- SI breakdown. Investigations of additional popula- len stainability than the parental accessions (Table tions throughout its range are needed to ascertain 3). Nearly all of the F1 hybrid plants had both the predominant breeding system in this taxon. unbranched and branched pubescence, and most From an ecological perspective, SI is common had both 4-merous and 5-merous flowers on the in Neotropical trees from lowland rain forest (Bawa same plant. Although I have only done limited 1979, Bawa, Perry, & Beach 1985), semi-decidu- studies of herbarium material, specimens have been ous forest (Bawa 1974), and deciduous forest (Ruiz seen that conform in morphology to the synthe- Zapata & Arroyo 1978, Bullock 1985). At La Sel- sized F1s(e.g., Donnell Smith 6672, the type of C. va, Kress and Beach (1994) reported that 88 per- Crossing Studies in Witheringia 77 cent of the hermaphrodite tree species that have or they may result from distinct allopatric entities been studied are SI. Fewer studies are available on that have come back into contact secondarily. breeding systems of tropical herbs and shrubs, but The studies reported here involved a small SC appears to be more common than in tropical number of collections of the W. solanacea complex trees (Schemske 1981, Whalen & Anderson 1981, from a limited part of its range. Whether or not Kress 1983, McDade 1985, Renner 1989). Kress these results can be extrapolated to the W. solanacea and Beach (1994) emphasized the disparity be- complex over its entire distribution is unclear, but tween predominant breeding systems in understory future taxonomic conclusions should take the fol- versus canopy and subcanopy taxa at La Selva, with lowing points into account: (1) all three entities are 34 percent SI species in the understory and 84 per- morphologically distinguishable, although the dif- cent SI species in the upper forest strata. They cited ferences between W. asterotricha and W. solanacea this pattern as evidence supporting the vertical are slight and may become blurred in areas of pos- stratification of breeding systems in a fashion sim- sible hybridization; (2) all accessions of W. astero- ilar to the vertical stratification of pollination tricha and W. solanacea tested were completely in- mechanisms described by Bawa, Bullock et al. terfertile in greenhouse crosses; and (3) W. meian- (1985). The abundance of SI species in the upper tha was genetically isolated from the other two forest strata as compared to the understory may be taxa. Perhaps the most rational scheme for dealing a real pattern, or it may prove to be a sampling with these issues is to recognize W. solanacea and artifact. Most of the SI understory species at La W. meiantha as distinct taxa, and to subsume W. Selva (10 out of 13) belong to a single family, the asterotricha within W. solanacea. Such taxonomic Rubiaceae (Bawa & Beach 1983, Kress & Beach decisions must await a complete systematic study 1994). The proportion of SI understory species at of the complex throughout its range. La Selva may change radically when more taxa are For convenience, a key to the three taxa of the investigated, especially in families in which SI is complex is included below. This includes observa- known to be common (e.g., Solanaceae, Rubiaceae, tions of the plants used in the study as well as her- Commelinaceae, Poaceae, Asteraceae, Melastoma- barium specimens from Duke University and Costa taceae; Fryxell 1957, Charlesworth 1985, Renner Rica, but is based on Costa Rican material only. 1989). Two solanaceous species now may be added to the list of SI understory taxa at La Selva, raising Key to the Witheringia solanacea complex the total percentage to 37.5. in Costa Rica A. Plants glabrous; flowers in fascicles of 10 (–15) or fewer; plants of forest light gaps ...... W. meiantha INTERTAXON CROSSES.—Genetic isolation appears to A. Plants pubescent; flowers in fascicles of generally be nearly complete between W. meiantha and both more than 10; pioneer plants of disturbed areas W. asterotricha and W. solanacea. Failure of crosses and secondary vegetation ...... B between these taxa cannot be explained by differ- B. Pubescence of dendritically branched hairs; flowers pentamerous; fruits pubes- ences in chromosome number or by low pollen cent ...... W. asterotricha fertility of the parents. Crossing barriers take the B. Pubescence of unbranched or rarely form of pollen tube growth inhibition, early abor- forked hairs; flowers often tetramerous; tion of hybrid seed, and nonviable seed formation. fruits glabrous or very sparsely pubescent ...... W. solanacea In contrast, hybrids were produced easily be- tween W. asterotricha and W. solanacea. Although pollen stainability of the hybrid plants was some- ACKNOWLEDGMENTS what lower than that of the parents, the F1 plants I thank C. LeBlanc and C. Christensen for help with green- were vegetatively vigorous and flowered profusely. house and laboratory work, D. Kropf and J. Sperry for use There is some evidence from limited herbarium of laboratory facilities, E. Dean, J. Denslow, and B. Loiselle and field studies that natural hybridization takes for access to unpublished information, M. Sousa-Pen˜a, B. Hammel, and L. McDade for comments on an earlier ver- place when pollen exchange between the two taxa sion of the manuscript, A. Torti and O. Lo´pez for correcting can occur. These two taxa are examples of semi- the Spanish abstract, the herbaria at F and US for loans of species (sensu Grant 1963, Mayr 1963, Grant specimens, the greenhouse staff at the University of Utah for 1981), which are morphologically or geographically maintaining my living collections, and the Organization for Tropical Studies and INBio for facilitating fieldwork in Costa differentiated entities that retain the ability to ex- Rica. Financial assistance to C. LeBlanc and C. Christensen change genes. These may represent species in the was provided by grants from the University of Utah Under- early stages of genetic isolation and differentiation, graduate Research Initiative. 78 Bohs

LITERATURE CITED

BAWA, K. S. 1974. Breeding systems of tree species of a lowland tropical community. Evolution 28: 85–92. . 1979. Breeding systems of trees in a tropical wet forest. N.Z. J. Bot. 17: 521–524. , AND J. H. BEACH. 1983. Self-incompatibility systems in the Rubiaceae of a tropical lowland wet forest. Am. J. Bot. 70: 1281–1288. ,S.H.BULLOCK,D.R.PERRY,R.E.COVILLE, AND M. H. GRAYUM. 1985. Reproductive biology of tropical lowland rain forest trees. II. Pollination systems. Am. J. Bot. 72: 346–356. ,D.R.PERRY, AND J. H. BEACH. 1985. Reproductive biology of tropical lowland rain forest trees. I. Sexual systems and incompatibility mechanisms. Am. J. Bot. 72: 331–345. BLAKE,J.G.,AND B. A. LOISELLE. 1992. Fruits in the diets of Neotropical migrant birds in Costa Rica. Biotropica 24: 200–210. BOHS, L. 1991. Crossing studies in Cyphomandra (Solanaceae) and their systematic and evolutionary significance. Am. J. Bot. 78: 1683–1693. BULLOCK, S. H. 1985. Breeding systems in the flora of a tropical deciduous forest in Mexico. Biotropica 17: 287– 301. CHARLESWORTH, D. 1985. Distribution of dioecy and self-incompatibility in angiosperms. In P. J. Greenwood, P. H. Harvey, and M. Slatkin (Eds.). Evolution: essays in honour of John Maynard Smith, pp. 237–268. Cambridge University Press, Cambridge, England. D’ARCY, W. G. 1973. Solanaceae. In R. E. Woodson and R. W. Schery (Eds.). Flora of Panama, pp. 573–780. Ann. Mo. Bot. Gard. Vol. 60. St. Louis, Missouri. . 1991. The Solanaceae since 1976, with a review of its biogeography. In J. G. Hawkes, R. N. Lester, M. Nee, and N. Estrada-R. (Eds.). Solanaceae III: Taxonomy, chemistry, evolution, pp. 75–137. Royal Botanic Gardens, Kew, England. FRYXELL, P. A. 1957. Mode of reproduction of higher plants. Bot. Rev. 23: 135–233. GRANT, V. 1963. The origin of adaptations. Columbia University Press, New York, New York. . 1981. Plant speciation, 2nd edition. Columbia University Press, New York, New York. HAUSER,E.J.P.,AND J. H. MORRISON. 1964. The cytochemical reduction of nitro blue tetrazolium as an index of pollen viability. Am. J. Bot. 51: 748–752. HEISER, C. B., JR., AND P. G. SMITH. 1958. New species of Capsicum from South America. Brittonia 10: 194–201. HUNZIKER, A. T. 1969. Estudios sobre Solanaceae. V. Contribucio´n al conocimiento de Capsicum y generos afines (Witheringia, Acnistus, Athenaea, etc.). Primera parte. Kurtziana 5: 101–179. . 1979. South American Solanaceae: a synoptic survey. In J. G. Hawkes, R. N. Lester, and A. D. Skelding (Eds.). The biology and taxonomy of the Solanaceae, pp. 49–85. Academic Press, London, England. . 1984. Estudios sobre Solanaceae. XIX. Sinopsis de Vassobia. Kurtziana 17: 91–118. . 1987. Studies on Solanaceae. XXI. A preliminary synopsis of Cuatresia. Opera Botanica 92: 73–82. JAIN, S. K. 1976. The evolution of inbreeding in plants. Annu. Rev. Ecol. Syst. 7: 469–495. KRESS, W. J. 1983. Self-incompatibility in Central American Heliconia. Evolution 37: 735–744. , AND J. H. BEACH. 1994. Flowering plant reproductive systems. In L. A. McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn (Eds.). La Selva: ecology and natural history of a Neotropical rain forest, pp. 161–182. University of Chicago Press, Chicago, Illinois. LOCKWOOD, T. E. 1973a. A taxonomic revision of Brugmansia (Solanaceae). Ph.D. dissertation, Harvard University, Cambridge, Massachusetts. . 1973b. Generic recognition of Brugmansia. Bot. Mus. Leaflets 23: 273–284. LOISELLE,B.A.,AND J. G. BLAKE. 1990. Diets of understory fruit-eating birds in Costa Rica: seasonality and resource abundance. Stud. Avian Biol. 13: 91–103. MARTIN, F. W. 1959. Staining and observing pollen tubes in the style by means of fluorescence. Stain Technol. 34: 125–128. MAYR, E. 1963. Animal species and evolution. Harvard University Press, Cambridge, Massachusetts. MCDADE, L. A. 1985. Breeding systems of Central American Aphelandra (Acanthaceae). Am. J. Bot. 72: 1515–1521. ,K.S.BAWA,H.A.HESPENHEIDE, AND G. S. HARTSHORN. 1994. La Selva: ecology and natural history of a Neotropical rain forest. University of Chicago Press, Chicago, Illinois. MURRAY, K. G. 1987. Selection for optimal fruit-crop size in bird-dispersed plants. Am. Nat. 129: 18–31. . 1988. Avian seed dispersal of three Neotropical gap-dependent plants. Ecol. Monogr. 58: 271–298. ,S.RUSSELL,C.M.PICONE,K.WINNETT-MURRAY,W.SHERWOOD, AND M. L. KUHLMANN. 1994. Fruit laxatives and seed passage rates in frugivores: consequences for plant reproductive success. Ecology 75: 989–994. NEE, M. 1986. Solanaceae. I. Flora of Veracruz. Vol. 49. Instituto Nacional de Investigaciones sobre Recursos Bio´ticos, Xalapa, Veracruz, Mexico. NETTANCOURT, D., DE. 1977. Incompatibility in angiosperms. Springer-Verlag, Berlin, Germany. OLMSTEAD,R.G.,AND J. D. PALMER. 1992. A chloroplast DNA phylogeny of the Solanaceae: subfamilial relationships and character evolution. Ann. Mo. Bot. Gard. 79: 346–360. ,J.A.SWEERE,R.E.SPANGLER,L.BOHS, AND J. D. PALMER. 2000. Phylogeny and provisional classification of the Solanaceae based on chloroplast DNA. In M. Nee, D. E. Syman, R. N. Lester, and J. P. Jessop (Eds.). Solanaceae IV: Advances in biology and utilzation, pp. 111–137. Royal Botanic Gardens, Kew, England. Crossing Studies in Witheringia 79

PLOWMAN, T. C. 1973. The South American species of Brunfelsia (Solanaceae). Ph.D. dissertation, Harvard University, Cambridge, Masssachusetts. PREIßEL, U., AND H.-G. PREIßEL. 1991. Brugmansia (Datura): Engelstrompeten. Ulmer, Stuttgart, Germany. RENNER, S. S. 1989. A survey of reproductive biology in Neotropical Melastomataceae and Memecyclaceae. Ann. Mo. Bot. Gard. 76: 496–518. RUIZ ZAPATA,T.,AND M. T. K. ARROYO. 1978. Plant reproductive ecology of a secondary deciduous tropical forest in Venezuela. Biotropica 10: 221–230. SCHEMSKE, D. W. 1981. Floral convergence and pollinator sharing in two bee-pollinated tropical herbs. Ecology 62: 946–954. VAN DER PIJL, L.. 1982. Principles of dispersal in higher plants. Springer-Verlag, Berlin, Germany. WELLER, S. G., M. J. DONOGHUE, AND D. CHARLESWORTH. 1995. The evolution of self-incompatibility in flowering plants: a phylogenetic approach. In P. C. Hoch and A. G. Stephenson (Eds.). Experimental and molecular approaches to plant biosystematics, pp. 355–382. Monographs in Systematic Botany Vol. 53. Missouri Bo- tanical Garden, St. Louis, Missouri. WHALEN,M.D.,AND G. J. ANDERSON. 1981. Distribution of gametophytic self-incompatibility and infrageneric classification in Solanum. Taxon 30: 761–767. WHEELWRIGHT, N. T., W. A. HABER,K.G.MURRAY, AND C. GUINDON. 1984. Tropical fruit-eating birds and their food plants: a survey of a Costa Rican lower montane forest. Biotropica 16: 173–192.